2025 (almost) in review: For the insurance industry, the future got here early

In November 2024, SAS and the Economist published a briefing that examined four possible futures for the insurance industry, highlighting the obstacles, the opportunities, and especially the global events that may shape the world of 2040.

While not intended to predict outcomes, the scenario analysis centered around two dynamics: global cooperation and pace of technological change. As world events unfold, we are already seeing several of the paths described in the report come true:

• “The US and China implemented restrictive policies on cybersecurity, intellectual property (IP) and trade.”
• “Isolationist policies slowed down innovation and economic growth.”
• “State-backed insurance programs struggled to fill the insurance coverage gap.”

That report projected 15 years ahead. Less than a year later, the future has arrived ahead of schedule.

The devastating impacts of the Palisades and Eaton fires “wiped out” the Fair Plans reserves and reinsurance, resulting in a $2 billion assessment, which then prompted a lawsuit filed by Consumer Watchdog. About $100 billion of the $162 billion in economic losses were in fact insured (due to most of the damage being isolated to the US).

American tariff policies have thrown the global economy into chaos. One study conducted by the American Property Casualty Insurance Association (APCIA) sites about 60% of new vehicle parts and 50% of replacement parts come from Canada, Mexico and China. They estimate the impact of a 25% tariff on Canada and Mexico and a 10% increase on China would result in a $7 billion to $24 billion increase in claims costs alone (note the resulting tariffs on China are significantly higher).  

The rushed deportation of hundreds of people to El Salvador has been well-covered, including its impact to those individuals’ families. While the human cost of these actions may not be quantifiable, the wider economic impact of such an isolationist policy could reduce the US’ GDP by 2.6% to 6.2%. The study further suggests that the role of immigrants in shoring up labor shortages would come under threat. Mass deportations “would likely lead to higher costs, increased inflation, and slower economic growth,” especially in states like California, Texas, and Florida (which are already experiencing severe insurance capacity challenges due to climate risk).

Given these events (and likely more to come), both trust and affordability will erode faster than many experts already anticipate. Only about half of US consumers trust their insurer, and rate increases post-COVID are at decades-high double-digits.

In a global survey of over 500 insurance executives from Economist Impact (sponsored by SAS), respondents rallied around a “bright spot” amidst the daily onslaught of gut-wrenching news: “78% of insurance executives believe closing the $1.8 trillion protection gap is an ethical obligation.”

The obstacles noted in that survey include (wait for it) diminished trust and lack of affordability.

So what’s next?

Regardless of the day-to-day drama of US politics, it’s clear that turmoil will continue. Insurers must address increasing instability for individuals and businesses. To the extent that technological innovation, new product development and resiliency investments can stabilize risk, insurers indeed have a pivotal moment to make a meaningful difference in the lives of billions of people.

And we need the insurance industry at the table. According to a US Senate Budget Committee report, Next to Fall: The Climate-Driven Insurance Crisis Is Here - and Getting Worse, a looming financial crisis to rival both 2008 and COVID has the potential to set off a chain reaction, devastating the global economy. Insurers must take a proactive look ahead and prepare should our society meet such a future.

What’s AI got to do with it?

The cost to train GPT-4, the large language model powering OpenAI’s ChatGPT in 2023, is estimated to be over $100 million (as compared to its predecessor GPT-3 at $2 million to $4 million a mere three years earlier). Last June, Time reported the power required to train AI models has doubled about every nine months since 2016, and could surpass $1 billion for some AI systems as early as 2027. 

Faster AI chips could ultimately reduce cloud computing costs. However, a single chip is already estimated to cost about $40,000, and the key ingredients to constructing them include rare earth elements, the majority of which come from countries like China. And data center spending is estimated to reach $1 trillion by 2031 (with the AI portion hovering in the mid 20% range for CAGR).

What does this mean for the 92% of insurers who are budgeting for generative AI projects in 2025?

The costs of those projects will increase. The $1 billion price tag for AI systems will come as soon as this year. First-time final model production and deployment must become status quo to protecting policyholders’ hard-earned premium dollars.

And as former Google CEO Eric Schmidt said during congressional testimony in April, the energy gap needed to power AI remains a growing concern. With 67 additional gigawatts of capacity needed by 2030 (insert Doc Brown joke here), he calls for broad investments in infrastructures and all forms of energy to bridge the gap.

So as costs and demand increase, how are insurance executives thinking about AI?

In the SAS and Economist survey I referenced, insurance executives viewed AI and market dynamics with cautious optimism, meaning they believe those trends a) will shape the sector in the coming decade and b) represent opportunities rather than risks. About half of those same executives also listed “minimizing underwriting and operational costs” as a key outcome for taking advantage of industry trends.

For example, rendering an immediate decision (for, say, a new policy application or claim) can and does create a lasting and positive customer experience; it also significantly reduces the expense burden of lengthy decisions and long processes. Both outcomes positively affect ROI.

At a recent customer event hosted by SAS and the Economist in Toronto, one executive shared that their organization is constantly focused on providing value to the policyholder through technological investments.

Artificial intelligence uses involving new policy acquisition, customer onboarding and claim settlements not only exist, they are being deployed by the largest insurers on the planet, at scale.

Will that matter if the cost of insurance coverage to the consumer or business represents an unaffordable expenditure?

Probably not. Outcomes from AI investments must be compelling, with tangible benefits. If insurers invest tens of millions of dollars for AI capabilities without delivering value that customers can see, those customers will feel betrayed.

How to right the ship?

If I’ve painted a dark picture of our current path – well, that was the idea. But failure isn’t inevitable. Next week, we’ll follow up with what should be a brighter look ahead – at how the insurance industry can still use AI and other technologies to steer away from those darker outcomes and toward a sunnier future.